Portland State University PDXScholar

Biology Faculty Publications and Presentations Biology

11-2001 High Frequency of Extra-Pair Paternity in Eastern Diane L. Rowe

Michael T. Murphy Portland State University, [email protected]

Robert C. Fleischer

Paul G. Wolf

Follow this and additional works at: http://pdxscholar.library.pdx.edu/bio_fac Part of the Biology Commons, and the Ornithology Commons

Recommended Citation Rowe, D. L., M. T. Murphy, R. C. Fleishcer, and P. G. Wolf. 2001. High frequency of extra-pair paternity in Eastern Kingbirds. Condor 103:845-851.

This Article is brought to you for free and open access. It has been accepted for inclusion in Biology Faculty Publications and Presentations by an authorized administrator of PDXScholar. For more information, please contact [email protected]. SHORT COMMUNICATIONS 845

The Condor 103:845±851 ᭧ The Cooper Ornithological Society 2001

HIGH FREQUENCY OF EXTRA-PAIR PATERNITY IN EASTERN KINGBIRDS

DIANE L. ROWE,1,4 MICHAEL T. M URPHY,2,5 ROBERT C. FLEISCHER3 AND PAUL G. WOLF1 1Department of Biology, Utah State University, Logan, UT 84322 2Department of Biology, P.O. Box 751, Portland State University, Portland, OR 97207 3Molecular Genetics Laboratory, National Zoological Park, Washington, DC 20008

Abstract. Genetic parentage in the socially monog- York central. En el 60% (12 de 20) de los nidos se amous and territorial Eastern (Tyrannus tyr- identi®caron juveniles con origen extra-pareja. De los annus) was examined in a central New York popula- 64 pichones investigados, el 42% fue engendrado por tion by multilocus DNA ®ngerprinting. Extra-pair machos fuera de la pareja, aunque no se detectaron young were identi®ed in 60% (12 of 20) of nests. Of nidadas con parasitismo conespecõ®co. Estos resulta- the 64 nestlings pro®led, 42% were sired by extra-pair dos son considerablemente diferentes a los obtenidos males, but no cases of conspeci®c brood parasitism en un estudio previo para la misma especie en una were detected. These results are markedly different poblacioÂn de Michigan, el cual reporto que el 39% de from a previous electrophoretic study of the same spe- los pichones no estaban relacionados con uno (tipica- cies in a Michigan population, which reported 39% of mente a la madre, cuasiparasitismo) o ambos (parasi- nestlings were unrelated to one (typically the mother, tismo de nido conespecõ®co) padres putativos. En la quasiparasitism) or both (conspeci®c brood parasitism) poblacioÂn de Nueva York, la paternidad extra-pareja of the putative parents. In the New York population, fue maÂs comuÂn entre hembras que retornaron a criar a extra-pair paternity was most common among females territorios que habõÂan ocupado previamente. Entre las that returned to breed on a former territory. Among hembras que ocuparon por primera vez un territorio de females that were new to a breeding territory, extra- crõÂa, la paternidad extra-pareja aumento directamente pair paternity increased directly with breeding density. con la densidad de individuos reproductivos. A pesar Although the power of the tests was low, neither que el poder del anaÂlisis fue bajo, ni la sincronõÂa re- breeding synchrony nor male experience with a breed- productiva, ni la experiencia de los machos en sus te- ing territory appeared to be associated with the occur- rritorios de cria, parecen estar asociados a la ocurrencia rence of extra-pair young. de juveniles extra-pareja. Key words: DNA ®ngerprinting, Eastern Kingbird, extra-pair fertilization, parentage, Tyrannus tyrannus. Genetic monogamy is now accepted as the exception rather than the rule among socially monogamous pas- Alta Frecuencia de Paternidad Extra-Pareja en serines that breed outside of the tropics. Westneat and Tyrannus tyrannus Sherman (1997) showed, for instance, that extra-pair young (EPY) were found in 86% of species Resumen. Se examino la paternidad geneÂtica de for which data were available (n ϭ 49 species), and Tyrannus tyrannus, especie socialmente monoÂgama y that on average, 18% of young were unrelated to one territorial, mediante la teÂcnica de huellas dactilares ge- of the parents, usually the male. Although a ®nding of neÂticas de muÂltiples loci en una poblacioÂn de Nueva extra-pair paternity is no longer surprising, it is still unclear why extra-pair fertilizations (EPFs) occur so ubiquitously. Adaptive scenarios based on female con- Manuscript received 8 April 2000; accepted 10 July trol of EPFs include the possibility that females ac- 2001. quire good genes, produce a more genetically diverse 4 Present address: Department of Wildlife & Fish- brood, gain insurance against male infertility, or obtain eries Sciences, Texas A&M University, College Sta- material bene®ts from extra-pair males. tion, TX 77843-2258. Questions also remain as to why the frequency of 5 Corresponding author. E-mail: [email protected] EPFs varies so widely among species. Two potentially 846 SHORT COMMUNICATIONS important factors are breeding density (Westneat and temperature for several months, but unfortunately, the Sherman 1997) and breeding synchrony (Stutchbury DNA from nearly all the nestling samples was severely and Morton 1995). Frequent contact among individuals degraded (adult DNA was ®ne; see also Conrad et al. breeding at high density may increase the potential for 2000) and no usable samples resulted from the 1994 EPFs (e.g., Reyer et al. 1997), and according to Stutch- season. Thus, in 1995 and 1996 we kept blood on ice bury and Morton (1995), so should high breeding syn- while we were in the ®eld, and after transport to the chrony if females control and seek extra-pair copula- laboratory, we stored all samples at 4ЊC until the DNA tions. In fact, Stutchbury and Morton (1995) proposed was isolated (2±3 days later). Usable ®ngerprint pro- that the degree of overlap in female fertile periods was ®les were generated for 20 families, 41 adults, and 64 the primary contributor to interspeci®c differences in nestlings. The odd number of adults is attributed to a EPF frequency. They hypothesized that synchronously nest at which three adults were captured (two males breeding species should have the highest EPF frequen- and one female). All young hatched from clutches of cies because females can simultaneously compare 2(n ϭ 1),3(n ϭ 11), or 4 (n ϭ 8) eggs, but three of many males and better select extra-pair mates. Much the nests having four young yielded samples for only of the available data support their hypothesis (e.g., three nestlings because of either the loss of a nestling Stutchbury et al. 1998, Chuang et al. 1999), but more to predators or insuf®cient DNA. data are needed to test the breeding synchrony hy- Total genomic DNA was isolated as in Fleischer et pothesis, and to test for an impact of other factors such al. (1994). Multilocus DNA ®ngerprinting was per- as breeding density, parental age, and breeder experi- formed using the Jeffreys' 33.15 probe (Jeffreys et al. ence on the occurrence of EPFs. 1985) and an M13 bacteriophage DNA probe for a The Eastern Kingbird (Tyrannus tyrannus) is a so- subset of seven families, four in which EPFs were and cially monogamous passerine that has been reported to three in which EPFs were not detected (Vassart et al. have a high frequency of extra-pair parentage. Curi- 1987). Standard ®ngerprinting methods were followed ously, females were usually excluded as the parent of (see Loew and Fleischer 1996 for protocols), with the the nearly 40% of young that were identi®ed as extra- exception of labeling M13 probe with [␥32P] dATP in- pair in a Michigan population (McKitrick 1990). As stead of [␥32P] dCTP, and then precipitating with salt, part of a larger study investigating the evolution of ethanol, and yeast tRNA. The DNA ®ngerprints for reproductive behavior in kingbirds, we used multilocus each family, arranged with nestlings ¯anked by their DNA ®ngerprinting to determine parentage and mea- putative parents, were manually scored on acetate sure the frequency of lost parentage among both males sheet overlays (Westneat 1990), and the number of and females in a New York population. We also at- novel bands and bands shared with one or both parents tempted to determine if nest density, nesting synchro- counted. ny, or parental experience with a territory were asso- STATISTICAL ANALYSES ciated with extra-pair parentage. We calculated band-sharing coef®cients (S of Lynch METHODS 1990) between nestlings and putative parents to re- solve parentage. S was calculated as twice the number Fieldwork was conducted in Delaware (42Њ78ЈN, of shared bands divided by the total number of bands 74Њ53ЈW) and Otsego (42Њ28ЈN, 75Њ03ЈW) Counties in for both individuals. The expected level of S for ®rst- central New York, where a color-banded population of order relatives was estimated by assuming that an adult Eastern Kingbirds has been under study since 1989 was the unambiguous parent of a nestling if there were (Murphy 1996, 2000). Nests were located prior to egg no novel bands in the nestling's ®ngerprint (0.55; Table laying by censusing all potential and former territories. 1). The expected S for ®rst-order relatives (R ϭ 0.50) We used adult behavior (territorial defense and feeding was calculated from the background level of S and of nestlings) to identify putative parents at individual equation 22 of Lynch (1990) as 0.56, very close to the nests. Nearly half of the adults upon which our work mean values of S for females and for unexcluded males was based were banded in a previous year, and none with offspring (Table 1). To determine the lower limit changed mates during the study. Mate replacement is of S below which ®rst-order relatives would not be a very rare phenomenon in this population (only one expected, we calculated the 95% and 99% con®dence case in 11 years among marked ); therefore we intervals surrounding the average S for ®rst-order rel- feel con®dent that all of the birds that were unbanded atives (0.55). These resulted in conservative lower lim- at the start of the study remained with their partners its of 0.38 and 0.32, respectively (Burke and Bruford through the entire nest cycle. 1987). Adults were captured using mist nets when nestlings We did not know the ages of the adults because none were 11±14 days of age. Birds were weighed and mea- were banded as nestlings. Instead, we categorized sured, and blood samples (Ͻ100 ␮L) taken via brachial adults as either ``experienced'' or ``inexperienced'' venipuncture. Unmarked birds were banded with one based on prior use of a speci®c territory. An experi- aluminum U.S. Fish and Wildlife Service band and a enced was a banded adult that bred on a territory unique combination of three color bands. Nestlings in one year and returned to breed on that same territory were weighed, measured, banded, and bled (50±100 in the next year. The designation of a bird as inexpe- ␮L, brachial venipuncture) when 13 days of age. In all rienced meant that it was unbanded and new to a ter- cases blood samples were collected in heparinized cap- ritory. These latter birds were known to be new to a illary tubes and immediately suspended in 1000 ␮Lof territory because they replaced a banded bird that had lysis buffer (Longmire et al. 1988). Blood samples col- bred on the territory in the previous season. An inex- lected in 1994 (n ϭ 20 families) were stored at room perienced bird may have bred elsewhere in the past, SHORT COMMUNICATIONS 847

TABLE 1. Band-sharing coef®cients (S) for presumptive father-offspring, presumptive mother-offspring, and unrelated adults (mated pairs) compared by the number of novel fragments. Results from the Jeffreys' 33.15 probe and M13 probe are summarized in (A) and (B), respectively. Values are mean S Ϯ SD (n), and ranges.

Number of novel fragments Relationship 00±2Ն3 (A) Jeffreys 33.15 Presumptive father-offspring 0.55 Ϯ 0.09 (24) 0.54 Ϯ 0.09 (37) 0.12 Ϯ 0.08 (27) Range 0.35±0.70 0.35±0.78 0.00±0.29 Presumptive mother-offspring 0.54 Ϯ 0.10 (24) 0.54 Ϯ 0.10 (37) 0.56 Ϯ 0.10 (27) Range 0.37±0.69 0.33±0.76 0.40±0.70 Male-female pair Ð Ð 0.12 Ϯ 0.09 (26) Range 0.00±0.30 (B) M13 0±1 Ն2 Presumptive father-offspring 0.48 Ϯ 0.14 (13) 0.16 Ϯ 0.12 (7) Range 0.33±0.70 0.00±0.30 Presumptive mother-offspring 0.61 Ϯ 0.14 (13) 0.45 Ϯ 0.15 (7) Range 0.36±0.86 0.24±0.64 Male-female pair Ð 0.11 Ϯ 0.12 (9) Range 0.00±0.33 but we suspect that most inexperienced birds were we estimated mutation rate per individual (m)tobe probably ®rst-time breeders because most kingbirds 0.36 (Westneat 1990). The probability of a nestling show high site ®delity between years (Murphy 1996). having three bands due to mutation was thus 4.7 ϫ We also evaluated the potential impact of breeding 10Ϫ2 (0.363; ϭ 3 of 64 young), which suggested that synchrony on extra-pair paternity (EPP) by calculating three novel bands could be used as a cutoff to identify Kempenaers' (1993) breeding synchrony index for all EPY. Thus, our ®nal criteria for using the Jeffreys' females in the population. The index was calculated as 33.15 probe to identify EPY were that nestlings have the percentage of other females in the population three or more novel fragments and an S below 0.32 whose fertile period overlapped that of the focal fe- with one or both parents. The M13 probe revealed few- male, where the fertile period was assumed to extend er fragments than the Jeffreys' 33.15, with an average from ®ve days before the laying of the ®rst egg until of 9.7 Ϯ 2.3 bands scored per individual. The mutation the laying of the penultimate egg of the clutch. We rate, m, was 0.23, so we used a cutoff of two novel also calculated the local breeding synchrony index for bands to identify EPY. all females for which we had ®ngerprints. The latter Based on Jeffreys' 33.15, EPY were detected in 12 index was based on the fertile periods of the four fe- of 20 nests (60%), and involved 27 of 64 nestlings males breeding closest to the focal female. We used (42%; Fig. 1). The mean S between presumptive fe- the average distance to neighboring pairs as an index male parents and nestlings did not vary with the num- of breeding density, by measuring the shortest distance ber of novel bands (Table 1), and averaged 0.55 Ϯ 0.10 from a focal nest to nests of the same four pairs that (n ϭ 64). Conspeci®c brood parasitism was thus never were used to calculate local breeding synchrony. All detected, and the probability of including a nonrelative nest locations were mapped on U.S. Geological Survey as the female parent was 9.9 ϫ 10Ϫ5 (Burke et al. topographic maps (1:24 000) for another study before 1989). On the other hand, mean S between the pre- the ®ngerprinting results were obtained. We used SAS sumptive father and young varied with the number of (SAS Institute 1990) and STATISTIX (Analytical Soft- novel bands, owing to the very low S of young with ware 1994) to test for relationships between the pres- three or more novel bands (Table 1). The average S of ence of EPY and breeding experience, nesting syn- the latter group of young with the presumptive father chrony, and density. Results are presented as means Ϯ was identical to that of unrelated adults in the popu- SD. Tests are described in the Results, and unless oth- lation (Table 1). The probability of inclusion of a non- erwise stated, we assumed signi®cance when P Յ 0.05. father was 2.1 ϫ 10Ϫ4 (Burke et al. 1989). EPP was RESULTS thus common and accounted for all cases of lost par- entage. DNA FINGERPRINTING RESULTS The probability of undetected instances of extra-pair The number of scorable bands in the 2±24 kbp range fertilizations was relatively low for this sample of 64 averaged 14.0 Ϯ 3.0 (range 8±21) for the Jeffreys' nestlings (1.3 ϫ 10Ϫ2). There was, however, one case 33.15 probe. The proportion of nestlings with different of ambiguous paternity. We captured two adult males numbers of novel bands was bimodal (24, 12, 1, 5, 4, and a female at a nest with three offspring in 1995. 8, 8, and 2 with 0, 1, 2, 3, 4, 5, 6, and 7 novel bands, The two males were determined to be ®rst-order rela- respectively), and the frequency of zero, one, or two tives (possibly brothers) based on a high S (0.64). Nei- 2 novel fragments ®t the Poisson distribution (␹ 1 ϭ 1.0, ther male could therefore be excluded as the true par- P Ͼ 0.5). Based on the ®t to the Poisson distribution, ent of the young. The primary male (netted closer to 848 SHORT COMMUNICATIONS

cording to the 33.15 probe, were reanalyzed with M13. The probability of assigning an unrelated male as fa- ther was 5.2 ϫ 10Ϫ5. Background band-sharing was also low (S ϭ 0.11 Ϯ 0.12), and we calculated the lower limit of the 95% and 99% con®dence intervals for male-offspring relatives as 0.29 and 0.22, respec- tively. Exclusions were clearly supported for 7 nes- tlings excluded by the Jeffreys' 33.15 data (Ն3 novel bands, S Ͻ 0.29; Table 1). For three nestlings, each with only a single novel M13 fragment, two had S of 0.46 and 0.40, while one had an S of 0.15. This third individual was not excluded based on the Jeffreys' 33.15 probe (1 novel fragment, S ϭ 0.46). The 10 re- maining nestlings were not excluded based on M13 data (as we also concluded using the Jeffreys' 33.15 probe). Thus, results from the two probes differed slightly for only 1 of 20 nestlings, and we are con®dent of the exclusions made using the 33.15 data alone. INDIVIDUAL AND ECOLOGICAL CORRELATES OF EPP The distribution of EPY among nests appeared bimod- al. Eight broods contained no EPY, two broods (brood size ϭ 3 and 4) contained one EPY each, but 10 broods had at least half of the young fathered by a male other than the presumptive father (7 broods of 3 and 3 broods of 4). Three males in fact failed to father any of the young in their nest (all broods of three). Assum- ing all young had a probability of 0.42 of being EPY, we calculated the number of broods that would be ex- pected to have EPY following methods described in Lifjeld et al. (1993). The predicted number of broods with EPY was 16.4 (out of 20), which differed mar- 2 ginally from the observed value of 12 (␹ 1 ϭ 3.282, P ϭ 0.07; cell totals corrected for small sample size). In an attempt to explain this pattern, we tested four po- tential ecological correlates of extra-pair paternity (EPP): male and female breeding experience, breeding synchrony, and breeding density. All six females that had experience on their territory in the past year had EPY, compared to only 6 of 14 inexperienced females (2 ϫ 2 contingency table, Fis- her's exact test, P ϭ 0.04). In contrast, there was no FIGURE 1. Examination of the occurrence of extra- association of EPP with male experience (Fisher's ex- pair young in Eastern Kingbird nests in New York. act test, P ϭ 0.64). In addition, we compared the num- Number of novel fragments and band-sharing coef®- ber of nests with EPY between pairs that were above cients (S) of young with the presumptive mother (up- versus below the median breeding synchrony index. per frame) and presumptive father (lower frame) for On both a population and local level, the occurrence the Jeffreys' 33.15 probe data only. The dashed lines of EPP was independent of breeding synchrony (Fis- indicate cutoff points for extra-pair young (Ն3 novel her's exact test, P ϭ 0.65 for both tests), but there was bands and S Ͻ 0.32). Points in the lower right quadrant a tendency for the number of EPY to vary inversely represent extra-pair young. with average nearest neighbor distance (calculated as the average distance to the four nearest neighbors). Eight of 10 nests below the median nearest neighbor distance held EPY compared to 4 of 10 above the me- the nest and having slightly higher S with the off- dian nearest neighbor distance (Fisher's exact test, P spring) was assumed to be the true father and no nes- ϭ 0.17). A natural break in EPY frequency occurred tlings were considered EPY (S ϭ 0.52, 0.54, and 0.64). at a nearest neighbor distance of 1 km: 10 of 13 pairs The female parent and the apparent male parent from with an average nearest neighbor distance Ͻ0.9 km 1995 returned to breed on the same territory in 1996 yielded EPY compared to 2 of 7 nests with an average (the only pair to contribute more than one ®ngerprint) nearest neighbor distance greater than 1 km (Fisher's and we again sampled the family unit. Surprisingly, in exact test, P ϭ 0.06). this pair's second year, the male did not father any of As a ®nal test, we performed a Poisson regression the young (S ϭ 0.08, 0.09, and 0.15). to simultaneously examine the in¯uence of female and Seven broods, four of which had EPF young ac- male experience, breeding synchrony, and nearest SHORT COMMUNICATIONS 849 neighbor distance on the number of EPY in a nest. All found that a third bird occasionally attaches itself to a four variables were entered, but then removed in a pair, generally late in the nest cycle. Although not backward stepwise procedure until we were left with common, it has occurred at least once in roughly every only those that signi®cantly reduced the model's de- other year of our 12-year study. We assume that these viance. The model retained female experience (P ϭ birds failed in a nesting attempt elsewhere and redi- 0.02) and average nearest neighbor distance (P ϭ rected their parental care (Bragg 1968). McKitrick 0.02). Separate comparisons of the number of EPY to (1990) did not follow pairs throughout the nest cycle, nearest neighbor distance for experienced and inex- and it is possible that some of the presumptive mothers perienced females showed that nests of experienced that she collected were actually unpaired birds that as- females held EPY regardless of density (b ϭϪ0.001, sociated with the nesting pair. Some of the difference P ϭ 0.76), whereas EPY were less likely to be found might also be attributed to problems inherent to elec- in the nests of inexperienced females as the average trophoretic techniques of assessing parentage. For in- distance to neighbors increased (b ϭϪ0.10, P ϭ 0.01). stance, differential gene expression between nestlings Least-squares linear regression indicated that 32% (P and adults may confound the assignment of parentage ϭ 0.04, n ϭ 14) of the variation in the number of EPY (Smyth et al. 1993), or the inherent low resolution of could be accounted for by nearest neighbor distance electrophoresis may make it dif®cult to discern wheth- among the inexperienced females. er observed mismatches are due to conspeci®c brood parasitism or EPFs (Romagnano et al. 1989, Smyth et DISCUSSION al. 1993). It is also possible that the difference between Our results provide evidence of a very high frequency the New York and Michigan populations is real, and of extra-pair paternity (60% of nests and 42% of off- although we regard it as unlikely, the possibility of spring) in an Eastern Kingbird population from central such a major intraspeci®c difference warrants further New York. Our recorded rate of EPP ranks among the study. highest yet reported for a socially monogamous pas- CORRELATES OF EPP serine (Fleischer 1996, Westneat and Sherman 1997). Although many studies have shown that EPP is com- As noted above, mate replacement is a very unlikely mon among , the individual and ecological explanation for our results since over an 11-year period predictors of EPP remain enigmatic. High nesting den- we have only once found that a banded male was re- sity, because it presumably increases the frequency of placed during the nesting cycle. Furthermore, we did interactions between extra-pair individuals, has been not increase the frequency of EPP through our capture argued to be an important contributor to the occurrence efforts because adults were not handled until nearly of EPP, but Westneat and Sherman's (1997) interspe- the end of the nestling period. We are thus con®dent ci®c comparisons provide little support. Within spe- that the adults identi®ed as the putative parents were cies, high nesting density tends to be associated with together from the start of nesting and that they behaved frequent loss of paternity (Reyer et al. 1997, Westneat normally. In addition, stored sperm from copulations and Sherman 1997; but see Chuang et al. 1999). In our that might have occurred prior to pair formation is a study, the number of EPY was negatively correlated very unlikely explanation for the high rate of EPP be- with nearest neighbor distance among inexperienced cause egg laying generally occurs two, and often three, females, suggesting an important in¯uence of nesting weeks after pairs form (but see Oring et al. 1992). density on the occurrence of extra-pair copulations. On Given the importance of last-sperm precedence (Birk- the other hand, all experienced females obtained EPFs head and Mùller 1992), copulation prior to pairing is regardless of nearest neighbor distance. probably of little consequence to kingbirds. We thus The high frequency of EPP in kingbirds is consistent believe that the EPP that we documented is the result with Stutchbury and Morton's (1995) hypothesis that of normal extra-pair fertilizations. synchronously breeding species should exhibit the McKitrick (1990) also found, using protein electro- highest frequency of EPP. The synchrony indices for phoresis, a very high rate of extra-pair parentage in a 1995 (0.55) and 1996 (0.54) rank among the highest Michigan population of kingbirds (39%), but oddly, recorded for temperate-zone breeding species (see Ta- the female was usually excluded as the probable par- ble 1 of Stutchbury and Morton 1995), and as their ent. Her data suggested that the social mother lost par- hypothesis predicts, the frequency of EPP in kingbirds entage as a result of both quasiparasitism (an unrelated is also high. On the other hand, we have no evidence female breeds with the mate of a female and then lays that breeding synchrony affected the within-population in the latter's nest; Wrege and Emlen 1987) and con- probability of EPY, but we offer this only as a tentative speci®c brood parasitism. Quasiparasitism has only conclusion because of the low power of our tests due rarely been documented in other birds (Wrege and Em- to small sample size. len 1987, Birkhead et al. 1990, Alves and Bryant The strongest in¯uence on the frequency of EPP ap- 1998), and conspeci®c brood parasitism (Rohwer and peared to be female experience: all of the females in Freeman 1989) has repeatedly been shown to be much our sample that returned to breed on a territory that less common than EPP. Thus, our failure to exclude they had used in the previous year obtained EPFs, re- the social mother as the genetic mother of even a sin- gardless of their proximity to neighbors. Conversely, gle nestling in the Charlotte Valley kingbird population less than half of the inexperienced females had EPY leaves us with the dif®cult task of interpreting the con- in their broods. Most surprising was the fact that ex- tradictory results of McKitrick's (1990) study. One perienced males that were paired with a former mate possibility is misidenti®cation of the actual parents in (i.e., both partners bred on the same territory and with the Michigan population. As described above, we have each other) frequently lost paternity, even of entire 850 SHORT COMMUNICATIONS

broods. It is our view that experienced females, be- FLEISCHER, R. C. 1996. Application of molecular meth- cause they knew the locations of males in the sur- ods to the assessment of genetic mating systems rounding landscape from the previous year, were able in vertebrates, p. 133±161. In J. D. Ferraris and S. to obtain EPCs with little dif®culty. The alternative R. Palumbi [EDS.], Molecular zoology: advances, interpretation of our data, that experienced females strategies and protocols. Wiley-Liss, New York. were more likely to suffer unwanted EPCs than inex- FLEISCHER, R. C., C. L. TARR, AND T. K. PRATT. 1994. perienced females, seems most unlikely. We therefore Genetic structure and mating system in the Palila, propose that experienced female kingbirds sought and an endangered Hawaiian honeycreeper, as as- obtained copulations from extra-pair males. sessed by DNA ®ngerprinting. Molecular Ecology We thank colleagues at the Molecular Genetics Lab, 3:383±392. National Zoological Park, for advice, and The Friends JEFFREYS, A. J., V. WILSON, AND S. L. THEIN. 1985. of the National Zoo for providing funds to the Molec- Individual-speci®c ``®ngerprints'' of human DNA. ular Genetics Lab. MTM received ®nancial support for Nature 316:76±79. this project from the National Science Foundation KEMPENAERS, B. 1993. The use of a breeding synchro- (BSR 9106854), and from Hartwick College's Trust- ny index. Ornis Scandinavica 24:84. ees' Research Grants. Our research was further funded LIFJELD, J. T., P. O. DUNN,R.J.ROBERTSON, AND P. T. by grants to DLR from the American Museum of Nat- BOAG. 1993. Extra-pair paternity in monogamous ural History (Frank M. Chapman Memorial Fund), the Tree Swallows. Behaviour 45:213±229. Association for Field Ornithologists (Bergstrom LOEW, S., AND R. C. FLEISCHER. 1996. Multilocus DNA Award), Sigma Xi (Grants-In-Aid Of Research), and ®ngerprinting, p. 456±461. In J. D. Ferraris and the E. L. and Inez Waldron Biotechnology Endowment S. R. Palumbi [EDS.], Molecular zoology: advanc- Fund. Additional support was provided by Utah State es, strategies and protocols. Wiley-Liss, New University: E. Brodie, Jr. and P. Gerity provided travel York. funds, and G. Podgorski provided laboratory space and LONGMIRE, J. L., A. K. LEWIS,N.C.BROWN,J.M. advice. This project would not have been possible BUCKINGHAM,L.M.CLARK,M.D.JONES,L.J. without the generosity of the landowners that allowed MEINCKE,J.MEYNE,R.L.RATLIFF,F.A.RAY,R. us access to their properties in the Charlotte Valley and P. W AGNER, AND R. K. MOYZIS. 1988. Isolation Otsego County. Finally, we thank several anonymous and molecular characterization of a highly poly- reviewers for their constructive comments on previous morphic centromeric tandem repeat in the family drafts of this manuscript. Falconidae. Genomics 2:14±24. LYNCH, M. 1990. The similarity index and DNA ®n- LITERATURE CITED gerprinting. Molecular Biology and Evolution 7: 478±484. ALVES,M.A.S.,AND D. M. BRYANT. 1998. Brood parasitism in the Sand Martin, Riparia riparia: MCKITRICK, M. C. 1990. Genetic evidence for multiple Evidence for two parasitic strategies in a colonial parentage in Eastern Kingbirds (Tyrannus tyran- passerine. Animal Behaviour 56:1323±1331. nus). Behavioral Ecology and Sociobiology 26: 149±155. ANALYTICAL SOFTWARE. 1994. STATISTIX. Version 4.1. Analytical Software, Tallahassee, FL. MURPHY, M. T. 1996. Survivorship, breeding dispersal BIRKHEAD, T. R., T. BURKE,R.ZANN,F.M.HUNTER, and mate ®delity in Eastern Kingbirds. Condor 98: AND A. P. KRUPA. 1990. Extra-pair paternity and 82±92. intraspeci®c brood parasitism in wild Zebra Finch- MURPHY, M. T. 2000. Evolution of clutch size in the es Taenopygia guttata revealed by DNA ®nger- Eastern Kingbird: tests of alternative hypotheses. printing. Behavioral Ecology and Sociobiology Ecological Monographs 70:1±20. 27:315±324. ORING, L. W., R. C. FLEISCHER,J.M.REED, AND K. E. BIRKHEAD,T.R.,AND A. P. MùLLER. 1992. Sperm com- MARSDEN. 1992. Cuckoldry through stored sperm petition in birds: Evolutionary causes and conse- in the sequentially polyandrous Spotted Sandpiper. quences. Academic Press, London. Nature 59:631±633. BRAGG, M. B. 1968. Kingbird feeding Baltimore Ori- REYER, H.-U., K. BOLLMANN,A.R.SCHLAPFER,A. ole nestling. Auk 85:321. SCHYMAINDA, AND G. KLECACK. 1997. Ecological BURKE,T.,AND M. W. BRUFORD. 1987. DNA ®nger- determinants of extra-pair fertilizations and egg printing in birds. Nature 327:149±152. dumping in Alpine Water Pipits (Anthus spinolet- BURKE, T., N. B. DAVIES,M.W.BRUFORD, AND B. J. ta). Behavioral Ecology 8:534±543. HATCHWELL. 1989. Parental care and mating be- ROHWER, F. C., AND S. FREEMAN. 1989. The distribution haviour of polyandrous Dunnocks Prunella mo- of conspeci®c nest parasitism in birds. Canadian dularis related to paternity by DNA ®ngerprinting. Journal of Zoology 67:239±253. Nature 338:249±251. ROMAGNANO, L., T. R. MCGUIRE, AND H. W. POWER. CHUANG, H. C., M. S. WEBSTER, AND R. T. HOLMES. 1989. Pitfalls and improved technniques in avian 1999. Extra-pair paternity and local synchrony in parentage studies. Auk 106:129±136. the Black-throated Blue Warbler. Auk 116:726± SAS INSTITUTE INC. 1990. SAS/STAT user's guide, ver- 736. sion 6. 4th ed. SAS Institute Inc., Cary, NC. CONRAD, K. F., R. J. ROBERTSON, AND P. T. B OAG. 2000. SMYTH, A. P., B. K. ORR, AND R. C. FLEISCHER. 1993. Dif®culties storing and preserving tyrant ¯ycatch- Electrophoretic variants of egg white transferrin er blood samples used for genetic analyses. Con- indicate low rate of intraspeci®c brood parasitism dor 102:191±193. in colonial Cliff Swallows in the Sierra Nevada, SHORT COMMUNICATIONS 851

California. Behavioral Ecology and Sociobiology minisatellites in human and animal DNA. Science 32:79±84. 235:683±684. STUTCHBURY,B.J.,AND E. S. MORTON. 1995. The ef- WESTNEAT, D. F. 1990. Genetic parentage in Indigo fect of breeding synchrony on extra-pair mating Buntings: a study using DNA ®ngerprinting. Be- systems in songbirds. Behaviour 132:675±690. havioral Ecology and Sociobiology 27:67±76. WESTNEAT,D.F.,AND P. W. S HERMAN. 1997. Density STUTCHBURY, B. J., E. S. MORTON, AND W. H. PIPER. and extra-pair fertilizations in birds: a comparative 1998. Extra-pair mating system of a synchronous- analysis. Behavioral Ecology and Sociobiology ly breeding tropical songbird. Journal of Avian 41:205±215. Biology 29:72±78. WREGE, P. H., AND S. T. EMLEN. 1987. Biochemical VASSART, G., M. GEORGES,R.MONSIEUR,H.BROCAS, determination of parental uncertainty in White- A. S. LEQUARRE, AND D. CHRISTOPHE. 1987. A se- fronted Bee-Eaters. Behavioral Ecology and So- quence in M13 phage detects hypervariable ciobiology 20:153±160.